No description available.

No description available.

The availability of large tranches of data and its influence on traffic flow, make the problem of short-term traffic speed prediction very complex in nature. For more than 40 years, various statistical time series forecasting methods have been applied for traffic speed prediction, and in the last 20 years, machine learning-based methods have gained prevalence. However, more recently, recurrent neural network (RNN) based methods have emerged to show better results for traffic speed prediction\cite{tian2015pred, zhao2017lstm,fu2017usin,chen2016long,dai2017deep,dai2019deep, kanestrom2017traf,shao2016traf,jia2017traf}. As the interest in applying RNN models to the traffic speed predictions started to grow, we found some critical and important unanswered questions with respect to the application of such methods. From these open questions, as part of this research study, we studied the following three questions for multi-step-ahead short-term traffic speed predictions: - What is the impact of using the direct and recursive strategies on the accuracy of RNN models as compared to using the multi-input-multi-output (MIMO) strategy? - What is the impact of different aggregation intervals for the input data? - What is the impact of including additional variables such as volume, occupancy, time of day, day of week, and spatial location as represented by station or sensor id? Our study resulted in the following observations, conclusions, and recommendations: - We observed that GRU architecture based RNN models had better accuracy as compared to other architectures. Thus we recommend that modeling efforts start with GRU architecture. - We observed that combining direct strategy with RNN gives the same accuracy as MIMO (i.e. many-to-many RNN architecture), however, the time taken for building many-to-many RNN architecture is much less as compared to direct strategy. Thus we recommend avoiding the use of recursive or direct strategies and advise the use of many-to-many RNN architecture without combining any other strategies. - We observed that there was no significant difference in the accuracy of 5,10,15 minute aggregations and that adding additional variables does not necessarily always result in higher accuracy. In both cases, we observed that using autotuning with a Bayesian algorithm was able to reduce the error metrics to a smaller range for almost all the combinations of aggregations and multivariate features. Among different data aggregations and feature sets, we suggest converting these choices as hyper-parameters and let the Bayesian algorithm based hyper-parameter tuner select the best combinations for your dataset. With the above contributions, this dissertation seeks to give traffic engineers a better understanding of how to apply modern methods for multi-step-ahead short-term traffic speed predictions.

When a driver turns a steering-wheel, he or she normally expects the vehicle's steering system to communicate an equivalent amount of signal to the road-wheels. This relationship is linear and occurs regardless of the steering-wheel's position within its rotational travel. The linear steering paradigm in passenger vehicles has gone largely unchanged since mass production of passenger vehicles began in 1901. However, as more electronically-controlled steering systems appear in conjunction with development of autonomous steering functions in vehicles, an opportunity to advance the existing steering paradigms arises. The following framework takes a human-factors approach toward examining and evaluating alternative steering systems by using Modeling and Simulation methods to track and score human performance. Present conventional steering systems apply a linear relationship between the steering-wheel and the road wheels of a vehicle. The rotational travel of the steering-wheel is 900° and requires two-and-a-half revolutions to travel from end-stop to opposite end-stop. The experimental steering system modeled and employed in this study applies a dynamic curve response to the steering input within a shorter, 225° rotational travel. Accommodation variances, based on vehicle speed and steering-wheel rotational position and acceleration, moderate the apparent steering input to augment a more-practical, effective steering rate. This novel model follows a paradigm supporting the full range of steering-wheel actuation without necessitating hand repositioning or the removal of the driver's hands from the steering-wheel during steering maneuvers. In order to study human performance disparities between novel and conventional steering models, a custom simulator was constructed and programmed to render representative models in a test scenario. Twenty-seven males and twenty-seven females, ranging from the ages of eighteen to sixty-five were tested and scored using the driving simulator that presented two successive driving test vignettes: One vignette using conventional 900° steering with linear response and the other employing the augmented 225° multivariate, non-linear steering. The results from simulator testing suggest that both males and females perform better with the novel system, supporting the hypothesis that drivers of either gender perform better with a system augmented with 225° multivariate, non-linear steering than with a conventional steering system. Further analysis of the simulated-driving scores indicates performance parity between male and female participants, supporting the hypothesis positing no significant difference in driver performance between male and female drivers using the augmented steering system. Finally, composite data from written questionnaires support the hypothesis that drivers will prefer driving the augmented system over conventional steering. These collective findings support justification for testing and refining novel steering systems using Modeling and Simulation methods. As a product of this particular study, a tested and open-sourced simulation framework now exists such that researchers and automotive designers can develop, as well as evaluate their own steering-oriented products within a valid human-factors construct. The open-source nature of this framework implies a commonality by which otherwisedisparate research and development work can be associated. Extending this framework beyond basic investigation to reach applications requiring morespecialized parameters may even impact drivers having special needs. For example, steeringsystem functional characteristics could be comparatively optimized to accommodate individuals afflicted with upper-body deficits or limited use of either or both arms. Moreover, the combined human-factors and open-source approaches distinguish the products of this research as a common and extensible platform by which purposeful automotive-industry improvements can be realized—contrasted with arbitrary improvements that might be brought about predominantly to showcase technological advancements.

In Florida, a state with significant population growth, it is essential to understand how land use change and transportation interact. Understanding these interactions between land use and transport is useful for transportation planners and land use modelers. Hence, in this research, we attempt to quantify the impact of the transportation system on land use and vice-versa. The research, using high-resolution land use data from 2011 to 2019, builds a binary logistic regression model of land use change. The model accounts for various independent variables, including socio-demographic attributes, built environment characteristics, and transportation network variables. The data set covers GIS data for three Central Florida counties: Orange, Osceola, and Seminole. The model building exercise is conducted using a random sample of parcel-level records. The model results indicate that as the distance to the nearest road or central business district (CBD) or shopping centers from the parcel increases, the likelihood of land use change is reduced. Furthermore, increasing the number of cafeterias or supermarkets or restaurants in the area lessens the likelihood of land use change. On the other hand, as the distance to the nearest recreational area or park or hospital from the parcel increases the likelihood of land use change is likely to be higher. The model developed was also validated using a holdout sample analysis.

No description available.

No description available.

No description available.

No description available.

No description available.